§ 17.] THONETICS. LIQUIDS. 9 



the Skr. caturdasii-. Av. *brddyya-, O. rasrai, a brother's son. In this the original O. 

 form may have been *^asrai, the first sr becoming r for ease of pronunciation. Cf. 

 sHsfak for *§rustak in § 29, ic. On the other hand, as stated in § 29, ig, there are no 

 instances of the occurrence of this change of br to ?'. Skr. vrlhi-, 0. rickan, rice. In 

 this the V probably represents an original vowel. Cf. Greek opvl^a. I^ess easy to 

 explain are Skr. grat'nati, he ties, O. granyi, a knot (cf. § 29, ic), and Av. hrazaiti, 

 Skr. h'rajate, 0. hnisi , he gUtters. 



(5) In the case of original compounds of which r is the first member, we have 

 Skr. pitrvya-, Av. tuirya-, O. ta, a paternal uncle, where there has been syncope of r 

 or ry. Cf. P. tr\ Ks a rule Skr. rt becomes O. r (§ 16, i), and Eranian rt, rd, rd, n, 

 and rn, become / (§ 17, 3). 



(6) Syncope of r occurs in Av. ^/kar-^ 0. ka-yek, to do ; Av. \/grab-, g9r9wnaiti , p.p. 

 g9rdpta-, 0. gla-s-tak or ga-s-tak , to take away; Skr. nihsarati, O. nisi, he emerges. 

 This kind of syncope is very common in the Pisaca languages. 



(7) There is apocope of ;' in Av. pitar-, 0. pie, a father; Av. matar-, 0. maw", a 

 mother. These may be considered as nursery words. 



§ 16. r. 

 (i) r, which is rare except as a medial, represents an original rt in Skr. garta-, a 

 hole, O. yir-ek, to hide. 



(2) r exceptionally represents an original I or r in Prs. kalUy, 0. krdy, a crow. 

 ? cf. 0. krdg", a hyena, the derivation of which is unknown to me. In O. pari, a 

 foot, the ri is a diminutive termination borrowed from India. The original is Av. 

 pdBa-, Skr. pada-. 



(3) Nearly every word containing r in 0. is borrowed from P. The above are 

 the only instances that I have noted which do not appear to have been so borrowed. 

 The word tar", a tank, is derived from Skr. tata-, and is evidently borrowed from 

 India. 



§17- I- 

 {1) I = original initial Skr. /, Eranian r, in Skr. lihanti, Av. raezaite, they lick ; 

 O. x/las-, lick ; Skr. lagyate, 0. laga, he is applied ; P. rawd-H, 0. lup-^ek, to be 

 suckled; Skr. ^lud-, 0. \/lar~, to stir; O. Prs. rddiy ; Prs. tU-rd, Kurd, -ta, ra ; P. 

 dar ; 0. dal, to thee. 



(2) The change of ^ to / occurs only in words borrowed from P., such as 0. lewu 

 (P. lew"), a wolf ; 0., lawanai, P. lewanai, mad. 



(3) Medial / = original Eranian medial r in Av. ^grab-, gdrdwnditi, p.p. gdrppta-, 

 0. gla-s-tak or ga-s tak, gli, he takes away ; Eranian s/'^gard-, O. gall, he twists; Av. 

 ddrdta-^ cut, O. ^dil-, to reap; Av. mdrdta- , dead, O. ^mul-, to die ; Av. sargta-, 

 O. sal' , coldness ; Av. sardZa-, O. sal, a year ; Av. sukurdna-, O. sukal, a porcupine ; 

 Av. zivdd-, 0. zll, the heart ; Av. ^/x^^^-. O- x^al-ak, to eat, but x'm^, he eats- (see 

 § 15, lb). We have seen in § 15, ib that medial r also = original r. The reason 

 why r has remained unchanged in some words, and has apparently become / in others, 

 is that in every case in which r has become /, in O. Prs. and Av. it is invariably 



